Acoustic emission (AE) signals of tomato plants were collected with AE signal detecting system. The collected data were processed and analyzed by using the upper machine software. By statistical analysis on changing of AE signal count with tomato late blight disease stress level, it was concluded that the plants with AE signal would not occur in physiological cycle like healthy plants. In the prophase of blight disease, the AE signal frequency dramatically increased, but decreased slowly in the crop disease progress. The multi-peak phenomenon emerged. Then, noise source location could be identified with time difference between two sensors and changing frequency distribution between different signals.